1. Generally, the initial angular light distribution is a function of two arguments:
(i) polar angle "theta" varied from 0 to 180 degree, where 0 is direction normal to the epilayers from n-side to p-side;
(ii) azimuthal angle "phi" varied from 0 to 360 degree;
In RATRO, the initial light distribution is assumed to be independent on the azimuthal angle (axisymmetric). From SpeCLED, RATRO
imports the lateral distribution of the emission power density (i.e. the amount of light power emitted from each cell), but not
the angular distribution.

As for dependence regarding polar angle, there are three options you listed.

2. "Isotropic" option means isotropic in 2D angular space (solid angle). So mathematical dependence of the probability function
regarding the angle theta is sin(theta). In other words, the probability for a ray to have the polar angle in the range from theta
to theta + dtheta is sin(theta) * dtheta.

3. Similarly, the values specified in "table" option are also multiplied by sin(theta). In other words, the red plot right to the
table is a cross-section of the 2D angular distribution.

4. Lambert type distribution has no direct physical meaning for light-emitting structures. In my opinion, the reasonable options
are isotropic and table (please note that one can make a table, say, in Excel, then save it as a text file, then import into RATRO
- no need to input data manually). Of course, "Lambert" option is a quick way to check what will happen if the light is tended
towards normal direction...



5. I suggest to use isotropic distribution until user have clear understanding of what is the physical reasons for anisotropy.
Generally, the angular distribution is anisotropic, and the degree of anisotropy depends on the crystal symmetry, structure
orientation, and ratio of contributions of different types of holes into the total emission.